PhD candidate: Milan van Puffelen
Start: Fall 2025
Given a timetable, it might not be possible to create a feasible crew schedule in which all tasks are covered in times of crew shortage. In this case, cancelling tasks can offer a solution. However, typical crew scheduling problems and methods do not include the option of cancelling tasks. Hence, my research focusses on partially integrating crew scheduling and timetabling by allowing to cancel tasks of an ideal timetable in the crew scheduling problem. The ideal timetable can be interpreted as the set of tasks one would ideally operate, but which is optimistic with respect to available crew.
As a solution approach, I apply column generation to a modified crew scheduling model that incorporates task-cancellation constraints capturing key rolling stock requirements. The method is developed and tested on real-world data from NS, ensuring practical relevance for railway operators.
The model supports NS in making timetable and crew decisions under staff shortages, instead of relying on ad‑hoc manual adjustments. By jointly modelling task cancellations, crew regulations, and rolling stock constraints, infeasible or contradictory plans that only surface during operations are prevented. The approach quantifies trade‑offs between service level and resource usage and can also be relevant for other transport and workforce planning settings with tight resource constraints.
